Infection of CNS by HIV can lead to encephalitis that presents clinically as HIV- associated dementia and HIV-associated neurocognitive disorders (neuro-AIDS). Despite the advent of HAART, at the late stage of the disease more than 11% of HIV patients suffer from HIV-associated dementia. Current antiretroviral drugs are ineffective at treating viral infection in the brain, because they cannot freely diffuse across the blood-brain barrier (BBB). Therefore, HIV viral replication persists in the CNS and continues to augment the neuropathogenesis process. The development of antivirals which can cross the BBB is therefore a critical unmet medical need. We propose a new strategy to prevent and treat HIV CNS infection, through application of the results of our recent discoveries. We found that attachment of a cholesterol group (cholesterol tagging) to a peptide fusion inhibitor (FI) endows it with the ability to penetrate the brain. Use of peptide FI as HIV antivirals is a clinically validated strategy. However, the only FI in clinical use, (enfuvirtide, Fuzeon(R)), cannot penetrate the CNS. The use of cholesterol tagging for BBB penetration would be compounded by two additional advantages: (i) localization of the peptide in the lipid rafts where fusion activation occurs, resulting in dramatically increased antiviral potency, and (ii) binding to plasma lipoproteins, resulting in increased half-life in vivo. Accordingly, we recently described a cholesterol-tagged HIV peptide, which is the most potent FI reported to date, H100-fold more potent than enfuvirtide. We propose here to assess the potential of cholesterol-tagged FI to treat retroviral infection of the CNS, using feline Immunodeficiency Virus (FIV) as a unique and credible model for neuro-AIDS.